Assessing the carbon footprint of data transmission

Nowadays, several websites provide an estimation of the carbon footprint of online content and services (e.g. e-mails, video streaming, video conferencing, etc.). However, despite the increasing interest in understanding the environmental impact of the Internet, it is hard to find studies that do a calculation of the impact of IP traffic based on measurements and well-defined methodologies. To fill this gap, Ficher et al. conducted research to assess the carbon footprint of data transmission on a backbone network through a simplified LCA methodology.
The researchers set up measurements to assess the emissions related to the transmission of 1 GB of data between two sites, testing different conditions regarding the infrastructure density (one short segment in the “dense” region of Paris vs one 700 km-segment from Paris to Montpellier) and the network activity (peak day vs off-peak day). While the few existing studies in this field focus mainly on the energy consumption of the equipment required for the transmission, Ficher et al. also considered the GHG emissions due to the manufacturing of these devices (routers, OTN switches, WDM multiplexers), as well as of the NOC devices and the optical fiber.
Results showed apparently low emissions, with 1,5 gCO2eq per GB on the Paris-Montpellier segment. However, the outgoing traffic from the Montpellier node amounts to 50 TB on an average day, with related emissions corresponding to traveling about 390 km by car.
The study also provides useful insights to find solutions to minimize the carbon footprint of data transmission. The emission factor of the electricity consumed by the equipment has a strong influence on the environmental impact of the use phase (which was showed causing more than 59% of the overall carbon footprint of the transmission), while extending the lifetime of the network devices would have a positive environmental effect on the production phase. Furthermore, the study highlights the opportunity to work on flow schedules over time and on bandwidth sharing to reduce the GHG emissions of online services.

To know more about the research methodology and results, see the research paper by Ficher et. on HAL Open Science.